observability-llm-obs

Answers questions about LLM performance, token usage, measured costs, response quality, and agent workflows using telemetry ingested into Elastic.

When to use it

Use for questions about LLM monitoring, token spend, model latency, prompt or guardrail failures, and agent tool-call chains.

Give it a question about LLM or agent behavior; it returns a conclusion based on Elastic telemetry or evidence supplied in the question.